Feeling Our Robot World

Some of the most amazing things that we take for granted are microscopic, such as the air we breath or the soft touch of silk.  These “little things” make existence possible for us humans, but remain the biggest challenges for robots.

Earlier this month a team of scientists, led by Professor Gerald Loeb, at the University of Southern California’s Viterbi School of Engineering developed BioTac – an artificial finger sensor that is able to detect textures better than our phalanges. This sensor is designed to replicate the human finger’s ability to feel motion, temperature, and the friction of different materials. Its uncanning ability to tell the difference between plastic and Gorilla Glass illustrates how robotic limbs may one day be truly bionic.

BioTac does not only look like a real finger, its “skin” layers actually contain sensors to detect different surfaces.  The outermost skin is made of soft flexible material with its own fingerprints. Inside there is a “liquid-filled layer” that is wrapped around a “hard, bone-like core.” As the BioTac finger gently moves across a surface its “skin” feels the vibrations of the materials as it passes through the liquid layer to a hydrophone (or underwater microphone).  This microphone then sends the waveforms to BioTac’s brain that processes the materials through a “vibration library” to isolate the varying compounds by its surface friction identity. Today, BioTac can distinguish between 117 different materials, including stationery, fabric, and wood with a 95% accuracy rate.

BioTac’s technology could soon be implemented in a new generation of prosthetic limbs as its engineering team have left the university to launch a startup called, Syntouch LLC.  Syntouch’s mission is to produce “tactile sensors for mechatronic systems”  and appears to be off to a promising start as they are already being sold to other research facilities and manufacturing industrial companies.

I guess the only thing BioTac can’t do, is give an annoying driver the finger… but who knows maybe there’s a bot for that too.